Proposing Novel Multi Objective Workflow Scheduling Schemes in Fog Computing

Abstract

In recent years, The rapid growth of Internet of Things (IoT) devices and enhancements in Infor- newlinemation and communication technology (ICT) have led to the production of a substantial volume newlineof data by IoT applications. Various types of workflow applications such as scientific comput- newlineing, health-to-home surveillance, traffic control, and business processing can be deployed on newlineIoT Devices. The traditional cloud-computing environment was not designed to process such a newlinemassive amount of data in near real-time due to the geographical distance of data centers and newlinethe resource-constrained nature of IoT devices and faces several challenges such as network newlinelatency and energy consumption. To overcome the challenges, Fog computing has emerged as newlinean expansion to traditional cloud computing that offers services in closer proximity to the net- newlinework edge, termed a cloud-fog environment. However, Workflow applications are collections newlineof interdependent tasks with a pre-defined order of execution. Scheduling workflow applica- newlinetions in such a heterogeneous cloud-fog environment with multiple conflicting objectives such newlineas maximizing performance, minimizing incurred cost, reducing energy consumption, etc., is newlinean NP-hard problem, and unsolvable in polynomial time. Meta-heuristics solutions are the best newlinechoice to optimize the multi-objective problems in polynomial time. The challenge with these newlinesolutions is to make a fair balance between local and global search capabilities. newlineHowever, several multi-objective scheduling solutions have been proposed for cloud en- newlinevironments, but very few exist for cloud-fog environments. Even the majority of solutions newlineproposed for cloud fog focus solely on either makespan or cost or both, but neglecting energy newlineconsumption as an optimization objective. newline